Accounting for different pathways to retirement is important for designing effective pension reforms, which are necessary to guarantee the financial stability of the pension system. For example, raising the statutory retirement age or cutting early retirement benefits are unlikely to reverse trends in early retirement if individuals can retire through the disability or unemployment insurance programs instead. The goal of our project is to understand the (financial) incentives that induce individuals to retire early, taking into account all available retirement channels and other factors that may affect the participation decisions of near-elderly workers. The project will make use of individual administrative data, which contains detailed information regarding earnings and employment histories, covering all of Austria in the period 1972 to 2007. Given the similarity in characteristics of the social security system and the aging population structure, our results can be applied largely to other industrialized countries. The combination of detailed administrative data and policy variation from a series of national pension reforms presents us with a unique opportunity to learn about individuals'retirement decisions in ways that are impossible in a U.S. setting. Because many of the Austrian reforms mimic policy changes that are currently debated by policymakers around the world, studying their effects can yield important insights to guide social security, disability insurance and unemployment insurance reforms in several other countries, including the United States. PUBLIC HEALTH RELEVANCE: Accounting for different pathways to retirement is important for designing effective pension reforms, which are necessary to guarantee the financial stability of the pension system. The goal of our project is to understand the incentives that induce individuals to retire early, taking into account all available retirement channels, including disability insurance, unemployment insurance and partial retirement. We combine both reduced form and structural methods, using administrative data from the census of Austrians combined with quasi- experimental variation from a series of pension reforms, to address this question.